420 



BOTANICAL GAZETTE 



[DECEMBER 



plants of this family with the other 3 families of the Piperales. 

 In all three of the latter, the Saururaceae, Piperaceae, and Lacis- 

 temaceae, the flowers are hermaphrodite (except in Piper Betel 

 L. var. monoicum CDC. according to Johnson 10, p. 716) and 

 are arranged in distinct catkins. The occurrence of unisexual 

 flowers is generally regarded as a mark of specialization within 

 the family or order, and their presence in Hedyosmum indicates 

 that this is one of the more specialized Piperales, rather than a 



very primitive one. 



Again, the perianth of Hedyosmum is not of a primitive type. 

 In fact it is a very unusual one. In so far as the writer has reviewed 

 the literature on angiosperms in general, and the Piperales in par- 

 ticular, he has not discovered any type of floral envelope which 

 is at all closely similar to it in its origin or mature structure. 



Clarke, who was the first to describe this perianth and to 

 note its pores, does not discuss its origin, but refers to it as a calyx. 

 Solms mentions a structure which he denotes as a perigon trifid 

 at the apex. Baillon says the apex of the ovary bears 3 short, 

 thick, rounded wings whose morphological value is still uncer- 

 tain. Eichler speaks of a 3-lobed half or quite superior perigon. 



- 



Bentham and Hooker mention a tube of the perianth as grown 

 to the ovary. 



The endosperm, which is cellular from the outset of its develop- 

 ment, finds its parallel in this respect in Heckeria and Peperomia, 

 but differs from that of Piper (Johnson 8, 10), the only other 

 genus of Piperaceae whose development is known. Elsewhere 

 among dicotyledons the type of endosperm formation found in 

 Hedyosmum occurs only in highly specialized forms (Johnson 9). 

 The families in which this is found are placed by Engler in the 

 higher orders of the Archichlamydeae and Metachlamydeae. It 

 would seem, therefore, that the conclusion reached by Hof- 

 meister, Strasburger, and Hegelmaier, that the structure and 



development of the gametophyte of angiosperms are not a satis- 

 factory index of broader genetic relationships, finds support here 

 in this peculiarity of the structure and development of the perianth 

 and endosperm of Hedyosmum. 



